GSEA showed that Mrtf-a–dependent genes
( 29 ) were up-regulated in ArhGEF1- and CD97-
deficient cDC2s (Fig. 5E). Many of the Mrtf-a–
regulated genes encode for cytoskeletal proteins
( 27 , 28 ), and several such genes were up-
regulated in CD97-deficient cDC2s (Fig. 5F).
Hematopoietic cells lacking Mrtf-a (and the
related gene Mrtf-b) show poor migratory
responses ( 29 , 30 ). Thus, Mrtf-induced gene
expression may contribute to altered motility
of cDC2s deficient in the CD97 pathway.
CD97 pathway restrains splenic cDC2 migration
We next used intravital microscopy to observe
splenic cDC migration. We intercrossedBatf3–/–
mice that lack cDC1s (fig. S6, G and H) ( 31 ) with
CD11c–YFP (yellow fluorescent protein) reporter
mice andArhgef1–/–mice. Analysis of tissue
sections showed that the majority of CD11c-
YFP+cells in these mice were DCIR2+cDC2s,
and their density was reduced in mice lacking
ArhGEF1 (fig. S6G). Real-time intravital imag-
ingrevealedthatalthoughmostcDC2sshowed
little migratory activity, occasional cells were
detected that were actively migratory (Fig. 5G
and movie S2). Rare cells could also be detected
moving into large blood vessels (likely red pulp
sinuses), in some cases through active migra-
tion(Fig.5GandmovieS3).Inothercases,cells
appeared to already be detached, rounded up,
and in flow within the red pulp at the time of
reaching a large vessel (movie S4). Like intra-
vascular DCs in dermal lymphatics ( 32 ), intra-
vascular cDC2s often continued to migrate for
a period of time before being carried away into
the blood flow (Fig. 5G and movies S3 and S4).
Despite the overall reduction in cDC2s detectable
per frame (fig. S6I), there was an increased fre-
quency of migratory cDC2s in ArhGEF1-deficient
mice (Fig. 5H and movies S5 to S8). Moreover,
an increased frequency of cDC2s was observed
moving into large blood vessels and being
caught in flow (Fig. 5I and movies S5 to S8).
In accord with the imaging data, there was an
increased frequency of ArhGEF1-deficient CD11c-
YFP+cDC2s in the blood circulation (Fig. 5J).As another approach to test for cDC2 move-
ment from the spleen into the blood, we trans-
planted spleens fromGna13cKOor control mixed
chimeric mice into WT recipients. Eight to nine
hours after the surgery, we examined the recip-
ient blood for donor-derived B cells and cDC2s.
The frequency of B cells in the blood closely
matched the chimerism in the transplanted
spleen, whereas cDC2s in blood were signfi-
ciantly enriched forGna13cKOcDC2s (Fig. 5,
K to N, and fig. S6J). Thus, Ga 13 -deficient
cDC2s have altered migratory behavior that
results in their loss from the spleen into the
blood circulation.Deficiency in the CD97 pathway alters
DC function
To test the impact of deficiency in the CD97
pathway on DC activation and function, we
immunized mice with a model systemic anti-
gen, sheep red blood cells (SRBCs). Three hours
after immunization, there was a selective defect
in antigen acquisition by the CD97-deficientLiuet al.,Science 375 , eabi5965 (2022) 11 February 2022 6 of 13
ACD97cDC2RBCs
RecipientCd55 Cd55
Cd55 Cd55 Cd55Cd551612840MFIof CD97 (×103 )**** ****B4.53.01.50MFI (×103 ) of CD976.0PE PESaline αCD55****n.s.6.04.53.01.50MFI (×103 ) of CD97Saline αCD55PE PEn.s. n.s.C
T419G O/E ChimerasD129630MFI of CD97 (×103 )Cd55 Cd55n.s.*IVC HeartCD97cDC2
43210MFI of CD97 (×103 )Cd55 Splenocytes
Cd55 RBCs
Cd55 RBCs
Shakern.s.n.s.
****n.s.n.s.EFCD97cDC2
43210MFI of CD97 (×103 )T419G mutant
Cd55 RBCs
Shakern.s.n.s.
n.s.Fig. 4. CD55-mediated CD97 NTF extraction is dependent on shear stress.
(A) Representative (left) histogram and (right) geometric mean fluorescence
intensity (MFI) of surface CD97 expression on cDC2s in mice with blood
transfusion as indicated. Data are pooled from three independent experiments.
(B) MFI of surface CD97 on in vivo PE-labeled (PE+) and PE-nonlabeled (PE−)
cDC2s in WT mice that had been treated 8 hours earlier with antibody to CD55 or
saline. Data are pooled from two independent experiments. (C) Chimeras were
reconstituted withAdgre5−/−BM transduced with Adgre5 (1, 2, 4)ÐT419G
mutant, with Thy1.1 as a reporter. MFI of surface CD97 on in vivo PE-labeled
(PE+) and PE-nonlabeled (PE−) Thy1.1+cDC2s in chimeras that had been treated
8 hours earlier with antibody to CD55 or saline. One of two independent
experiments with similar results is shown. (D) MFI of surface CD97 expression on
transferredCd55−/−cDC2s in blood collected from heart and distal ligated IVC of
recipients as indicated (movie S1). Data are pooled from two independent
experiments. (EandF) Splenocytes from (E)Cd55−/−mice or (F) CD97 T419G
expressing BM chimeras were cocultured for 1 hour at RT withCd55−/−or
Cd55+/+RBCs, on a shaker or not. Representative (left) histogram and (right)
MFI of surface CD97 expression on cDC2s are shown. Data are pooled from two
independent experiments. In (A), (E), and (F), color coding of histograms is as
for the graphs, and gray lines indicate isotype control. In (A) to (C), (E), and
(F), each symbol indicates one mouse, and lines denote means. In (D), lines
connect data from the same animal. *P< 0.05; ***P< 0.001; ****P< 0.0001.RESEARCH | RESEARCH ARTICLE